Skype has given a detailed explanation of why its service collapsed for two days last week and it looks like Windows was the main culprit.
In a recent blog post, Lars Rabbe explains that on December 22nd a cluster of support servers responsible for off-line instant messaging became overloaded causing Windows clients to crash. Non Windows versions of the software including those for Mac and iPhone were not affected by the original problem.
Although the initial overload only affected 40 per cent of users, between 25 and 30 per cent of the public supernodes which distribute Skype traffic were also affected causing the global meltdown.
"Although Skype staff responded quickly to disable the overloaded servers and to eliminate client requests to them, a significant number of supernodes had already failed," said Rabbe. "A supernode is important to the P2P network because it takes on additional responsibilities compared to regular nodes, acting like a directory, supporting other Skype clients, helping to establish connections between them and creating local clusters typically of several hundred peer nodes per each supernode."
Increased pressure on the unaffected supernodes seems to have created a cascade effect leading to the massive disruption to the network.
Skype says it has identified the bug in the Windows client, will provide updates later in the week, and will consider offering 'automatic' updates in the future to prevent further outages.